For fun.. measure the voltage that the 2 chips are running at..
Just put the voltmeter across the connections?
Driver question
- Thread starter bblumberg
- Start date
Just put the voltmeter across the connections?
oreo57
Well-known member
Parallel at the driver output, dimming off..
opps..
OK, so new drivers came, installed them and no joy. White channels are still not behaving well. Putting voltmeter across the white channels - they are drawing 200V! ***? Now I remove one of the 3 DreamChips and get blinking of white channels at 48.5 V.
If I put 3 white channels on old driver - get 129V and no light. the other 3 channels work fine, power up, and dim as expected. 0V dim, 3V much brighter, 9V much brighter yet. RB channel (x 3 chips) draws 80.1 V at 0 V on the dimming circuit (with a 1K resistor in line), 92v at 3V and 109V at 9 V.
I am at a loss to understand what is going on here. I contacted Alan at AC-RC for his comments, but it seems like these chips will not have functional white channels. Not what I had in mind....
oreo57
Well-known member
The 200 or 129 V is what your meter will read if the circuit is "open". i.e whites not conducting..
The driver will ramp up voltage till Amp output is achieved.
W/ a meter this is never..
Thus it will ramp up to full output..
Sorry to hear this didn't solve the problem.
Best guess is one (or more) white channels has defective diodes..
Check the voltage w/ only one chip at a time on the white channel..
you actually can't judge voltage w/ a meter in pulsed dimming (assuming that regardless of the "type" of dimming output is still PWM-d.
You meter is most likely averaging i.e at 109V "full" your "metered voltage" drops to 54.5 at 50%
(109 1/2 the time 0 the other half = averaged 54.5)
Technically the voltage never really drops..Only 2 outputs:
voltage at drive current and
"zero" (may not be exactly zero but hey..)
Pretty sure you got bad chips..
The driver will ramp up voltage till Amp output is achieved.
W/ a meter this is never..
Thus it will ramp up to full output..
Sorry to hear this didn't solve the problem.
Best guess is one (or more) white channels has defective diodes..
Check the voltage w/ only one chip at a time on the white channel..
you actually can't judge voltage w/ a meter in pulsed dimming (assuming that regardless of the "type" of dimming output is still PWM-d.
You meter is most likely averaging i.e at 109V "full" your "metered voltage" drops to 54.5 at 50%
(109 1/2 the time 0 the other half = averaged 54.5)
Technically the voltage never really drops..Only 2 outputs:
voltage at drive current and
"zero" (may not be exactly zero but hey..)
Pretty sure you got bad chips..
Thanks for the explanation. I didn't notice any individual diodes not lighting when the chip was blinking, but this makes a lot of sense. As you can imagine, I am pretty discouraged with the Dream Chips at this point after investing so much time and effort (not to mention $) into the drivers, build and troubleshooting. I am strongly considering going back to square 1 and building a standard LED light like my other 2 using O2surplus drivers. I thought I was being clever with the DreamChips, but this doesn't seem like the case....
oreo57
Well-known member
When you found the orig. chips seemed to require a V(f) above "Normal" did indicate either 1)manuf. out of spec (which was the hope) or 2)damaged diodes..
LEd's are pretty simple critters..
Was hoping it was 1)
Seems to be 2)
At some point the white channel was probably overheated..though who knows..
Still not sur if it is one chip or all that is defective..
LEd's are pretty simple critters..
Was hoping it was 1)
Seems to be 2)
At some point the white channel was probably overheated..though who knows..
Still not sur if it is one chip or all that is defective..
Yeah, I am guessing defective, too. Or at least not very durable. In my hands, these chips have seen the same treatment on all channels. Therefore, while it is possible that I did something to the white channels, it seems more likely that there was something strange about the whites to begin with. It is perhaps suggestive that the 2 white channels are the outer 2 rows of the chip (on either side) whereas the blue and violet are in the middle. These have been on the provided VGA coolers the entire time, but who knows about the uniformity of heat distribution on these... I am still waiting to hear from Alan.
oreo57
Well-known member
Just to be clear I didn't assume you overheated them.. Just somewhere along the line something went bad..
not getting a fast response from the seller is not good..
As a side note, you are not the first person to struggle w/ these large voltage drivers.. Though good in concept, not very good from a practical DIY standard..
not getting a fast response from the seller is not good..
As a side note, you are not the first person to struggle w/ these large voltage drivers.. Though good in concept, not very good from a practical DIY standard..
I didn't take it that way, but my experience in the lab suggests that either something comes bad, or you mistakenly make it bad. I will be pleasantly surprised if the seller assumes responsibility and replaces the chip...
Ordinarily the seller is fairly quick so let's see.
Thanks for not saying "I told you so" since you were not in favor of the high voltage drivers from the beginning. Ironically, the drivers seem to work perfectly - it is the chip that is giving me headaches.
The problem about these is just the amout of diode troubleshooting you need to do. Picture long strings of 3W diodes.. like 50..
Makes sense. In my case, these are strings of 30 diodes, packed closely together 10 at a time.
As for individual chips rather than multichips - individual chips gives you much more control over colour and placement, is (IMO) easier to trouble shoot and easier & cheaper to replace failed components. I'd still go multichip (and did for my build) for how much quicker and easier the physical build and wiring is!
Tim
Tim
I need to cut the splices in each one of these and find a new driver to test the channels individually - I think I can do that if I remove one of the drivers from my nanocube. The HLG-80 worked with 2 chips, but 1 chip doesn't draw enough voltage for me to test them one-by-one.
I am struggling with this decision at the moment. I am tempted by the BlueAcro 20K pro chip that Theatrus has designed, but will probably need at least 4 of these and perhaps 6 for my 150G (60 x 24 x 24) SPS build. The alternative I am considering is to simply get a large Makers heat sink and put the individual LEDs on it as I did with my 75G tank. I have 52 (mostly Cree XT-E and XP-G2) LEDS on that tank on a 6 x 20" heat sink, although they are running at about 50% (RB) or less (white, violet). Scaling this up would be around 125-ish LEDs on a 48" heat sink or ~150 on a 60". No small task to build, but perhaps the safest in terms of troubleshooting.
Thanks for you help.
So on your recommendation and that of the vendor, I tested the white channels of each chip individually using 3 x 9v batteries in series. The results were "illuminating". One chip worked perfectly on both channels, all LEDs illuminated brightly. One chip showed very faint illumination of all LEDs in channel 1 and bright illumination of all chips in channel 5 (this is the chip that let channel 1 function when I removed it). The other showed very bright illumination of channel 1 but nothing on channel 5. It seems pretty clear that these chips are the problem, not the drivers (which I had already surmised by testing the drivers on the blue and violet channels). Let's see what the vendor has to say about this...
Hopefully they will agree to replace, but as you suggested it is very easy to suggest the DIYer must have broken them. And to be fair to the retailers, all too often the DIYer does break them, and, although most people are, not everyone is honest 
Not for a second suggesting you damaged these, just I'd hate to be a retailer (even more than I hate being a buyer and receiving broken goods!).
I know one retailer that got a power supply returned as faulty. When he opened it, it had salt crystals on the inside. It had obviously been for a dip in the buyers tank, but of course they forgot to mention it!
Good luck and hope they do accept them and replace
Tim
Not for a second suggesting you damaged these, just I'd hate to be a retailer (even more than I hate being a buyer and receiving broken goods!).I know one retailer that got a power supply returned as faulty. When he opened it, it had salt crystals on the inside. It had obviously been for a dip in the buyers tank, but of course they forgot to mention it!
Good luck and hope they do accept them and replace
Tim
oreo57
Well-known member
If it is any consolation you only lose 13.3% of your light potential.
and if you run 1-1-5 say on a channel and save the "odd" leftover for the center.. it would be a nice moonlight..
Given lemons make lemonade..
Oh and at "the least" they should give a good faith discount if they won't do a full refund/replace.
and if you run 1-1-5 say on a channel and save the "odd" leftover for the center.. it would be a nice moonlight..
Given lemons make lemonade..
Oh and at "the least" they should give a good faith discount if they won't do a full refund/replace.
Last edited:
oreo57
Well-known member
Question is did he replace it?
